Abstract: Testing machine learning sensors by adding obfuscated training data to test data, and performing real time model fit analysis on live network traffic to determine whether to retrain.
Type:
Application
Filed:
July 23, 2020
Publication date:
November 19, 2020
Applicant:
Resurgo, LLC
Inventors:
Eamon Hirata Jordan, Chad Kumao Takahashi, Ryan Susumu Ito
Abstract: Testing machine learning sensors by adding obfuscated training data to test data, and performing real time model fit analysis on live network traffic to determine whether to retrain.
Type:
Application
Filed:
July 31, 2020
Publication date:
November 12, 2020
Applicant:
Resurgo, LLC
Inventors:
Eamon Hirata Jordan, Chad Kumao Takahashi, Ryan Susumu Ito
Abstract: Testing machine learning sensors by adding obfuscated training data to test data, and performing real time model fit analysis on live network traffic to determine whether to retrain.
Type:
Grant
Filed:
December 8, 2016
Date of Patent:
August 4, 2020
Assignee:
RESURGO, LLC
Inventors:
Eamon Hirata Jordan, Chad Kumao Takahashi, Ryan Susumu Ito
Abstract: Testing machine learning sensors by adding obfuscated training data to test data, and performing real time model fit analysis on live network traffic to determine whether to retrain.
Type:
Application
Filed:
December 8, 2016
Publication date:
June 14, 2018
Applicant:
RESURGO, LLC
Inventors:
EAMON HIRATA JORDAN, CHAD KUMAO TAKAHASHI, RYAN SUSUMU ITO, MATTHEW DAVID-KRISTOFER TROGLIA
Abstract: Heterogeneous sensors simultaneously inspect network traffic for attacks. A signature-based sensor detects known attacks but has a blind spot, and a machine-learning based sensor that has been trained to detect attacks in the blind spot detects attacks that fail to conform to normal network traffic. False positive rates of the machine-learning based sensor are reduced by iterative testing using statistical techniques.
Type:
Application
Filed:
November 3, 2014
Publication date:
February 19, 2015
Applicant:
RESURGO, LLC
Inventors:
Eamon Hirata Jordan, Kevin Barry Jordan, Evan Joseph Kelly
Abstract: Heterogeneous sensors simultaneously inspect network traffic for attacks. A signature-based sensor detects known attacks but has a blind spot, and a machine-learning based sensor that has been trained to detect attacks in the blind spot detects attacks that fail to conform to normal network traffic. False positive rates of the machine-learning based sensor are reduced by iterative testing using statistical techniques.
Type:
Grant
Filed:
March 14, 2013
Date of Patent:
November 11, 2014
Assignee:
Resurgo, LLC
Inventors:
Eamon Hirata Jordan, Evan Joseph Kelly, Kevin Barry Jordan